Abstract: Thin bed is a key factor influencing GR logs, while in return the accuracy of the logs restricts the precision of thin- bed thickness identification. In this paper, a correcting model of GR logs is firstly adopted to improve the accuracy of GR logging values, and then on the basis of it a mathematical model of correcting thickness is established, which corrects thin - bed thickness with the corrected GR logging values. The software based on the correcting method is applied to thin - bed layers of more than 10 wells in a certain reservoir which can correct their GR logging values and identify the accurate thickness of each layer. The results prove that this method can increase the accuracy of log interpretation and calculation of factual thickness of thin - bed reservoir.